Reproduce “Simulation-Free Learning of GP-SDEs from Irregular Observations” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Zhidi Lin, Yuhao Liu, Ying Li, Edwin Fong, Petar Djurić
GP-SDE Matching learns a Bayesian model of continuous-time dynamics from noisy observations arriving at irregular times. It represents the unknown drift with a Gaussian process and estimates state distributions at queried times, avoiding simulated trajectories in its dynamics loss. In a stochastic Lorenz–63 test, time-aware encoding reduced drift RMSE from 2.83 to 0.32 under sparse random-window observations.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.